Is data module may be applied (connected) only to relational packages?

[brain1e55]

Hey,

May I ask to clarify my understanding of how data modules work?
My initial guess was that I easily can supplement an existing DMR package with some additional info from a data module. But it turned out that I cannot connect an excel file to any dimensional item of a DMR package.

Does it mean that data modules are only useful when I have a strictly relational package?

Thanks.

[MFGF]

Hey,

May I ask to clarify my understanding of how data modules work?
My initial guess was that I easily can supplement an existing DMR package with some additional info from a data module. But it turned out that I cannot connect an excel file to any dimensional item of a DMR package.

Does it mean that data modules are only useful when I have a strictly relational package?

Thanks.

Hi,

If you have a dimensional package, the query engine will generate MDX rather than SQL, and MDX doesn't support a relational join paradigm. One option is to create a data set from the relevant data pulled from your dimensional package, and use the data set in your data module (you can schedule data sets to refresh automatically). Alternatively you could use the underlying tables feeding your DMR package and bring those into your data module.

Data Modules support uploaded files, data sets, relational packages and data servers (databases) being joined.

Cheers!

MF.

[brain1e55]

Thanks, MFGF.

Neither of these approaches allows me to supply additional info to the existing DMR package as If it would be a pure relational package, right? Moreover, I cannot do the same with a dynamic cube as it is being queried via MDX as well.
Meaning that if I am a regular CA user and I'd like to combine existing DMR package data and my own data (plan data, for example, to do the plan-fact analysis), I cannot easily do that in order to compare them.
What I can do according to your suggestion is to take out a relevant data set from the DMR package and make a data module on that base, adding then an excel file containing the plan data.
Actually, I'm getting more and more sceptical about this dimensional approach, since the benefits of data modules marketed by IBM Cognos have a very limited usage. And since many users want to plug their custom data in (self-service BI, bla-bla-bla), I need to find arguments in favor of the dimensional approach.


Shortcut: either do an extract from a DMR/DynCube and make a data module then, or work solely with data modules without invoking any existing OLAP-like packages.